Financebotresearch desk研究台

EMR › analyze

EMR · Analyze from before

WATCH Infrastructure

Date: 2026-08-27 | Price: $157.71 | 52wk: $122.64–$166.35 | Market Cap: $88.33B Baseline: analyze-2026-03-21.md (+ same-day health and sentiment files), price $128.15, verdict WATCH, conviction 7.0

1. What this updates

The March 21 baseline called EMR a strong, transformed business trading at base-case fair value ($125–140) with a target buy-the-dip entry of $105–118. The stock did not dip. It ran, gaining 23% to $157.71 while the S&P was up single digits, on the back of two consecutive beat-and-raise quarters. The entry zone on the watchlist ($105–118) is now 34% below spot and has been silently suppressing the name from every scan since — the exact failure mode in [[pitfall-stale-entry-zone-suppresses-a-name]]. This pass exists to re-derive the zone and test whether the price move was earned by the business or is running ahead of it.

Event list since March 21: - Q2 FY26 earnings (~May 2026) — details folded into quarterly statement data below. - Q3 FY26 earnings, reported Aug 4–5, 2026 — the dominant event. Revenue $4.873B (+7% underlying), adjusted EPS $1.71 (+13% YoY), FCF $1.323B (+36% YoY). Full-year guide raised: adjusted EPS to ~$6.55 (from $6.40–6.50), FCF to ~$3.6B. - A wave of analyst price-target raises and one upgrade (JPMorgan, Neutral→Overweight, Jul 17) following the print — reversing the "deteriorating" analyst trend the baseline flagged. - New Equinor (13-year) and BP automation contracts announced Aug 2026 — incremental order-book evidence, not thesis-changing on their own. - A new risk not present at baseline: Middle East conflict / Strait of Hormuz disruption, ~$100M annualized FY26 revenue headwind, first flagged on the Q3 call. - No change to the debt structure's trajectory in the direction baseline expected — see the ledger below.

2. The delta ledger

Ledgered against the test "would changing this claim change the recommendation." Structural and Trend rows drive thesis persistence; State rows are the current facts underneath them; Price and Judgment rows are re-derived, not carried.

Lead rows — retracted, superseded, drifted

# Claim (baseline) Status What changed / what carried it
18 "Forward EPS (FY2026) ~$7.15 consensus, forward P/E 17.9x" ❌ RETRACTED [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] fires on EMR. Yahoo's forwardEps was never FY2026 — it was already FY2027 consensus in March, and remains FY2027 today ($7.24 now). The true FY2026 figure was always the company's own guide, now confirmed at $6.55 (epsCurrentYear/priceEpsCurrentYear = $6.5554 / 24.06x, exactly matching management's raised guide). The baseline mislabeled a 12-months-further-out number as the current year, which is the same direction of error the pitfall always produces — it made the stock look cheaper than it was. See §5 for the pitfall-note amendment.
15 "Tariff exposure ~$130M built in, net neutral so far" 🔄 SUPERSEDED Tariffs partially reversed ($82M in cash refunds contributed to the Q3 FCF beat) while a new, unrelated geopolitical headwind emerged: Middle East conflict / Strait of Hormuz disruption, ~$25M Q3 impact, ~$100M annualized FY26 modeled impact, customer operating capacity recovering to ~75%. The specific threat rotated; the total risk budget is similar in size.
17 "Analyst consensus Buy but deteriorating" (Buy ratings 20→17 over 3mo, Oppenheimer + Deutsche Bank downgrades) 🔄 SUPERSEDED The trend reversed. Since baseline: JPMorgan upgraded Neutral→Overweight (Jul 17); price-target raises from Citigroup ($174→182), RBC ($169→179), Evercore ($185→190), Bernstein ($169→186), DA Davidson ($145→155, still Neutral). No new downgrades found this pass. Same mechanism (analysts tracking delivery), opposite direction, driven by two clean beat-and-raise quarters.
9 "FY2025 debt spiked to $13.8B total / $4.8B short-term... will come down as FCF grows and transformation spending normalizes" 📉 DRIFTED Two quarters later, debt has not come down. Total debt Q2 FY26 = $14.06B (up from $13.76B), short-term/current debt = $5.80B (up from $4.80B) — financed via an enlarged $7B commercial-paper program plus a $3B 364-day revolver backstop, and a $750M 0.875% note matures Oct 2026. This is not a distress signal (ample facility headroom, investment-grade issuer, well-telegraphed AspenTech financing) but the baseline's specific expectation of near-term deleveraging has not materialized. Breaks the thesis if net debt/EBITDA rises meaningfully past ~3x with no visible paydown path by FY2027, or a rating action follows.
6 "Revenue 3yr CAGR ~9.3%... FY2026 Net Sales Growth ~5.5%, Underlying ~4%" 📉 DRIFTED (mild) Both growth sub-guides were trimmed slightly at Q3: GAAP sales growth guide 5.5%→5%, underlying 4%→3.5% — the Middle East headwind and continued China softness account for most of it. Offsetting: the margin/EPS guide rose in the same call (adj EPS $6.40–6.50→$6.55, FCF unchanged-to-up). The deceleration is on the top line only; profitability is compensating. Not thesis-breaking, but the "9%+ organic CAGR" framing needs a haircut.

Carried and refreshed — compact list

# Claim Status Note
1 Transformation to pure-play automation/software complete (no HVAC/climate legacy) ✅ CARRIED Segment structure unchanged: Final Control, Measurement & Analytical, Discrete Automation, Safety & Productivity, Control Systems & Software, Test & Measurement.
2 DCS/process-control switching costs are a durable moat ✅ CARRIED Reinforced by new 13-year Equinor frame agreement and a BP automation contract — long-duration commitments are evidence of, not just an assertion of, switching costs.
3 AspenTech ~80%+ share in downstream refining simulation ✅ CARRIED No contradicting evidence found; Control Systems & Software segment (houses AspenTech) grew 7% at 31.8% margin in Q3.
4 Gross margin expansion trend (45.7%→52.8% FY22-25) 🔁 REFRESHED TTM gross margin now 53.16%, still expanding, though the pace of expansion has slowed from the +710bps/3yr run-rate.
5 FCF (cont. ops) 3yr CAGR ~22.8%, FY25 level $3.25B 🔁 REFRESHED True TTM FCF = $3.461B (Q4FY25 $842M + Q1FY26 $602M + Q2FY26 $694M + Q3FY26 $1,323M — summed, not read off one quarter; see §6 for the reconciliation, done specifically to avoid [[pitfall-single-quarter-fcf-read-as-ttm]]). FCF/share ~$6.18, up ~8% from baseline's $5.73. Q3 alone grew FCF +36% YoY. FY26 guide raised to ~$3.6B.
7 Shares outstanding declining ~1.5-1.7%/yr ✅ CARRIED 560.1M now vs 566.7M FY25 diluted. Buyback dollars down (~$898M 9mo FY26 vs $1.167B all of FY25) — debt financing is the capital-allocation priority right now, not accelerated repurchase — but the share count is still falling.
8 Dividend growth ~5%/yr recent, 1.9-2.2% 5yr/3yr CAGR 🔁 REFRESHED Nov 2025 hike: $0.528→$0.555/qtr = +5.1%, confirming the ~5% pace continues. Not the 7%+ the market needs for DDM to justify spot (see §6).
10 Net debt $11.6B / Debt-to-Assets 32.8% / Net debt/FCF ~3.6x ✅ CARRIED (tension noted) Net debt is flat ($11.57B FY25 → $11.57B Q2 FY26 — literally unchanged), but because TTM FCF grew, net debt/FCF improved mechanically to ~3.3x. Coverage is better even though the nominal debt figure the baseline worried about did not fall. Both things are true; they are not in conflict once you separate stock from flow.
11 Reported ROIC 6.7-9.9% below WACC, tangible-capital ROIC 37-54% ✅ CARRIED ROE now 12.8% (up from ~11.0%), ROA 6.9% — directionally consistent with baseline's goodwill-distortion read. Full tangible-capital ROIC recompute is UNTESTED this pass (see §7).
12 FCF payout ratio ~37% (cont. ops) ✅ CARRIED TTM FCF payout now ~36% ($2.22 ÷ $6.18) — dividend safety unchanged.
13 NI/Test & Measurement integration "too early to tell" (Moat) vs. "$200M synergies achieved" (Sentiment) 🔁 REFRESHED, resolved positive Test & Measurement grew +23% underlying in Q3 (up from +11% at baseline), Software & Systems overall +11% at 31.8% margin. Growth verticals +27%, led by semiconductor +53%, power +37% — NI is now visibly riding the AI-datacenter/grid-buildout tailwind, not just holding its own. The Moat Analyst's original "trust but verify" caution was reasonable; the near-term evidence has come in clearly positive.
14 China "turned a little more bearish," guided low-single-digit negative for 2026 ✅ CARRIED Q3 actual: China -3%, "sequential improvement" noted. Management's own forecast was accurate — a predicted risk that materialized exactly as sized is a force for management credibility, not against the thesis.
16 Insider activity: net selling, no buying on dips ✅ CARRIED Continued: CEO Karsanbhai sold Aug 11 ($161.91) and Dec 3 ($965,979 at $133); COO Krishnan sold Jun 1 ($141.22); Officer Train sold Aug 6 ($158.99). No purchases found. Same read as baseline: routine compensation management, not a confidence signal either way.
19 Fair value Bear $90-100 / Base $125-140 / Bull $170-190 🔁 REFRESHED Re-derived from scratch — see §6. New range: Bear $110-120 / Base $140-165 / Bull $175-195.
20 "At $128, EMR is trading at approximately base-case fair value" 🔄 SUPERSEDED Price has moved to $157.71 — now sitting at the top of the re-derived base case, not the middle. The margin of safety the baseline was waiting for has been consumed by the rally, not created by it.
21 DYT: yield 1.73% vs 5yr avg ~2.1-2.3%, "overvalued" 🔁 REFRESHED Same direction, wider gap: yield now 1.40% vs 5yr avg 2.00% (Yahoo fiveYearAvgDividendYield) — a larger relative shortfall than in March.
22 DDM fair value $67-159 across growth/discount scenarios 🔁 REFRESHED Same formula, same $2.22 dividend (unchanged since baseline). Spot has moved from needing the "optimistic" 7%-growth case to be justified ($118.77) to needing the "aggressive" 7.5%-growth case ($159.10) — while realized dividend growth is still running ~5%. The valuation stretch on this lens widened, it did not close.
23 Evergreen rating 7.5/10 ✅ CARRIED No erosion evidence; the Equinor/BP long-duration wins are mild reinforcement.
25 Key risk #1: Integration execution — HIGH 🔁 REFRESHED, downgraded to MEDIUM-LOW Two more clean quarters, synergies visibly delivering (row 13), two consecutive guide raises, no negative surprises. This was the baseline's top-ranked risk and it has substantially de-risked.
26 Key risk #2: Debt load / refinancing — MEDIUM-HIGH ✅ CARRIED Unresolved in the direction hoped (row 9) even as coverage improved (row 10). Still MEDIUM-HIGH; watch the Oct 2026 note maturity and year-end net debt print.

3. How the close calls were decided

Row 9 vs. row 10 (debt level vs. debt coverage) — the central tension this pass. The nominal debt figures moved against the baseline's stated expectation (total and short-term debt both higher, not lower), which on its own reads as a broken forecast. But the coverage ratio that actually determines serviceability — net debt against free cash flow — improved, because FCF grew faster than debt did. Neither row is wrong; they measure different things. The resolution: the balance sheet risk is real but has not worsened in the way that matters (coverage), even though it looks worse in the way that's easiest to headline (the dollar figure). Both rows are recorded separately rather than collapsed into one verdict, because a future pass that only checks the dollar figure would call this "deteriorating" when the debt-service math says otherwise.

Row 6 (growth guide trimmed) vs. row 13 (NI/growth-verticals accelerating) — do they contradict? No, once decomposed: the guide-down is concentrated in the denominator (GAAP/underlying total sales, dragged by Middle East disruption and continued China softness) while the acceleration is concentrated in a specific mix shift (semiconductor +53%, power +37%, Test & Measurement +23%) that is a smaller, faster-growing slice of the business gaining share of the whole. A slower-growing whole with a faster-growing, higher-quality core inside it is consistent with the margin guide rising in the same call that trimmed the sales guide — the company is being paid better for less volume, which is the software-mix thesis working, not failing.

Row 17 (analyst sentiment reversal) — is this thesis evidence, or just price chasing? Weighted as weak-to-moderate corroboration, not a primary force. Sell-side price targets moving after a beat-and-raise is expected and somewhat mechanical; it is recorded because it reverses a claim the baseline explicitly made, not because it independently proves anything. The stronger, primary evidence for the thesis is rows 4, 5, 13, and 14 — margin, cash flow, segment growth, and a correctly-forecast regional risk, all directly measured.

Row 18 (the forward-EPS mislabel) — does this change anything the baseline concluded? No. The baseline used the mislabeled $7.15/17.9x figure only as descriptive color in the header table; its actual fair-value work ran on FCF, DDM, and DYT, none of which touched the bad field. The error did not corrupt the March verdict. It is retracted here because it would have corrupted a trim level had one been set as a multiple against it, and because the same mechanism has now fired on this repo eight-plus times (TSM, QCOM, BSX×5, LEN, AMD, ANET) and belongs in the pitfall note's ticker list.

4. Thesis persistence and conviction delta

Thesis persistence: Structural claims (rows 1–3): 3/3 CARRIED. Trend claims (rows 4–8): 4/5 CARRIED or REFRESHED (row 6 DRIFTED). Persistence = 7/8 = 87.5%.

That is high persistence against a 23% price move — the textbook re-rating setup the command framework calls out: the business held (and in the case of NI/Test & Measurement, exceeded), and the multiple moved with it. This is the opposite of a thin thesis flipping on sentiment.

Conviction delta: 7.0 → 7.5. Driven by: - Up: Row 13 (NI/T&M integration risk resolved positive, +23% growth) and row 25 (the baseline's #1-ranked risk, integration execution, downgraded HIGH→MEDIUM-LOW) — together these remove the single largest uncertainty the March report carried. - Up (secondary): Row 14 (China risk materialized exactly as management forecast — a credibility point) and row 17 (analyst sentiment reversal, weighted lightly per §3). - Down (offsetting, keeps this at 7.5 and not 8+): Row 9 (debt has not delevered as expected), row 21/22 (DYT and DDM both show a wider valuation stretch than in March), and row 20 (the price has consumed the margin of safety rather than waiting inside it).

The net read: a better business than the market gave it credit for in March, now priced closer to what it's worth. Good news for the thesis, bad news for the entry.

5. What is genuinely new

  • Middle East / Strait of Hormuz disruption (row 15) — a geopolitical risk with no baseline counterpart, sized at ~$100M annualized FY26 revenue, customer capacity recovering to ~75%. Small relative to $18.9B guided FY26 revenue (~0.5%) but worth tracking for escalation.
  • Growth-vertical disclosure (semiconductor +53%, power +37%, part of row 13) — Emerson did not break this out in the March baseline's sources. It quantifies the AI-datacenter/grid-buildout exposure that the March sentiment report only described qualitatively ("NI... benefiting from secular demand for power infrastructure").
  • Pitfall-note update filed: [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] gains EMR as a confirmed case (see below) — the baseline's $7.15 "FY2026" figure was FY2027 consensus even in March, the same mechanism as TSM/QCOM/BSX/LEN/AMD/ANET.

6. Updated verdict

Valuation, re-derived from scratch

The fiscal-year check first, since it decides every multiple quoted below. EMR's fiscal year ends Sep 30 — FY2026 closes in ~5 weeks from this report's date. Yahoo's fields:

Field Value What it actually is
epsCurrentYear / priceEpsCurrentYear $6.5554 / 24.06x FY2026 — matches management's own raised guide (~$6.55) almost to the cent
forwardEps / forwardPE $7.237 / 21.79x FY2027 consensus — a genuinely "forward" year now, since FY2026 is nearly done
Trailing (GAAP, TTM) $4.57 / 34.51x Sum of last 4 reported quarters (1.12+1.07+1.10+1.28) — depressed by intangible amortization, not comparable across years

Unlike the baseline (where "forward" meant a year not yet begun), today FY2026 is nearly complete, so the current-year multiple (24.1x) is the more decision-relevant one for "is this cheap right now," and FY2027 (21.8x) is the honest "forward" figure for the trim convention below.

Model Fair Value / Signal Weight Note
Graham IV √(22.5×4.57×36.53) = $61.29 0% Still discarded — goodwill-heavy, buyback-hollowed book makes BVPS uninformative. (The baseline's own $17.50 BVPS input does not reconcile to either total book value [$36.53] or tangible book value [-$13.10]; immaterial since the model carried 0% weight both times, but flagged for the record.)
Bogle expected return ~11–12%/yr (1.40% yield + ~10.5% earnings growth ± modest multiple risk) 20% Down slightly from baseline's 11.7–13.8%, mainly the lower starting yield — still a respectable, not compelling, return.
DDM $46.62 (5%/10%) to $159.10 (7.5%/9%) 20% Spot now requires the aggressive growth case; realized growth is running ~5%. Signals rich.
DYT Yield 1.40% vs 2.00% 5yr avg — overvalued, wider gap than March 10% Consistent signal, strengthened.
FCF-based Bear $110-120 / Base $140-165 / Bull $175-195 50% Using the $3.6B FY26 FCF guide at 5.5%/4.0-4.5%/3.5% yields; 558M forward-adjusted shares.

Weighted fair value: $140-180. Spot ($157.71) sits inside this range, roughly at the midpoint to upper-third — fairly valued to modestly rich, not the clear-discount setup of March, but not a trim-now overvaluation either.

Entry / Trim (re-derived, not carried — the stale-zone rule requires this every pass)

  • Entry (accumulate): $135–148 — a discount to the base-case low-to-mid, roughly a 6–14% pullback from spot. This replaces the $105-118 zone, which was set against the March fair-value band and is now 34% below spot and un-reachable without a bear-case shock.
  • Strong buy / aggressive accumulate: below $122 — near the 52-week low and the bear-case FCF yield level.
  • Trim: 26x fwd. Basis: FY2027 consensus EPS $7.24 (the field the site's multiple-trim mechanism actually reads, per [[pitfall-stale-entry-zone-suppresses-a-name]]'s amendment on trim inheriting the vendor's forward-EPS field) → implied dollar ≈ $188. This sits above the current 21.8x FY2027 multiple, giving room for the "software-adjacent" re-rating (22-27x band, per the baseline's own tension-resolution table) to play out further before triggering, while still capping the position once the market prices in more than the 2028 targets currently justify. Check the number the site actually renders against this $188 figure at deploy time — if Yahoo's forward field moves materially before the next build, the rendered dollar will drift from this basis.

Verdict: WATCH — unchanged designation, better business, less margin of safety

EMR delivered on almost everything the March thesis needed and did so without a single negative surprise — NI/Test & Measurement is now clearly additive rather than a "trust but verify" bet, gross margin keeps expanding, FCF is compounding faster than the baseline modeled, and management called the one regional risk (China) correctly. That is why conviction moved up, not down. But the market noticed the same things, and the 23% rally has moved the stock from the bottom of its fair value range to the top of a higher fair value range — the DDM and DYT lenses both show a wider valuation stretch today than in March, and the debt load that was supposed to shrink has not.

This remains a name to own at a discount, not at the market. The business case has gotten stronger; the price case has gotten weaker in almost exact proportion. Re-set the zone, keep watching, and treat any pullback toward $135-148 as the entry the March baseline was originally waiting for.

Break triggers (replacing the March triggers, which never referenced a falsifiable event)

  • Net debt/EBITDA rises meaningfully past ~3x with no visible paydown path disclosed by the FY2026 close (Nov 4, 2026 print).
  • Test & Measurement or Software & Systems underlying growth decelerates below high-single-digits for two consecutive quarters (would undercut row 13, the biggest positive force this pass).
  • China underlying growth worse than the "low single digit negative, sequential improvement" path management has now correctly called twice.
  • A credit-rating action (downgrade or negative outlook) tied to the AspenTech-financing debt load.

Upgrade conditions (toward conviction 8+)

  • A pullback into the $135-148 entry zone with the thesis otherwise intact, or
  • Net debt visibly declining (not just coverage improving) by the FY2026 year-end print, confirming the deleveraging path the baseline assumed and this pass could not yet verify.

7. What this pass did NOT test

  • Full tangible-capital ROIC recompute (row 11). Directionally consistent (ROE up 11.0%→12.8%) but not independently re-derived from the current goodwill/intangibles split this pass — carried on the strength of the baseline's original decomposition plus the directional ROE move. Flag for the next full /analyze.
  • AspenTech's 80%+ simulation-software share (row 3). No new competitive data found either way; carried on absence of contradicting evidence, not fresh confirmation. This is now the second consecutive pass without a direct re-test — the next one should either re-verify it against a competitor filing/analyst note or flag it explicitly as stale.
  • Q3 FY26 balance sheet detail. The Yahoo MCP's quarterly balance sheet for the 2026-06-30 period returned almost entirely null fields; this report's debt figures use the most recent populated quarter (2026-03-31, Q2 FY26) plus qualitative confirmation from the Q3 earnings call. The exact Q3-end net debt number is UNTESTED — worth pulling from the 10-Q directly next pass (SEC EDGAR blocked WebFetch this session; a 10-Q filed via a different retrieval path should close this gap).
  • 2028 management targets ($21B revenue, 30% adj EBITA margin, $8.00 adj EPS, $12B cumulative FY25-28 FCF) — not re-tested against the trimmed FY26 sales guide. Worth a specific check next pass: does 5%→3.5% underlying growth in FY26 still pencil to $21B revenue by FY28, or does the target need a explicit reaffirmation from management.

Sources